Squeak-dev June 2003

squeak-dev@lists.squeakfoundation.org

188 participants
973 discussions

Re: [UPDATES] 7 for 3.6alpha
by Bruce ONeel 02 Jun '03

02 Jun '03

Thanks, these are now all packed up in: ftp://st.cs.uiuc.edu/pub/Smalltalk/Squeak/3.6alpha/Squeak3.6a-5247.zip cheers bruce Doug Way <dway(a)riskmetrics.com> wrote: > > These are the 4 items which were [approved] last week, plus one which > Daniel approved about a month ago that I somehow missed. (The subject > line for that one was munged a bit, though... I'll use that as an > excuse. ;-) ) > > - Doug Way > > > ------------------------------- > > 5241instrPrinter-md -- Marcus Denker -- 8 April 2003 > A refactoring for InstructionPrinter, > AbstractInstructionPrinter and InstVarRefLocator: > AbstractInstructionPrinter: > moved the instVar and a method into InstVarRefLocator, now this class > can > be used as a Abstract Superclass for > both InstructionPrinter and InstVarRefLocator > Renamed to InstructionClient (c.f. VisualWorks). > InstructionPrinter: refactored to be a > subclass of InstructionClient" > > 5242HashChanges1-SqR -- Andres Valloud, Stephan Rudlof -- 13 August 2002 > Goals: > Implement an improved hash for large integers. > Refactor/rename some hash methods to make them more intention revealing. > Pass 1 - Implement refactoring so it can be bridged later. > sr: I have changed #byteAt: to #basicAt: to make it Slang compilable > (and changed some var names)." > > 5243HashChanges2-SqR -- Andres Valloud -- 13 August 2002 > Pass 2. Implement LargePositiveInteger>>hash. > Rehash sets as appropriate." > > 5244HashChanges3-SqR -- Andres Valloud -- 13 August 2002 > Pass 3. Make String and ByteArray use the > refactoring instead of the original > implementation." > > 5245MorphicClsComments-efc -- Eddie Cottongim -- 7 March 2003 > Class Comments for Morph, ImageMorph, SimpleButtonMorph and > StringMorph. These were subject > to review and editing on the swiki, and thus should be fairly > comprehensive." > > 5246Sibling-dup-tk -- Ted Kaehler -- 11 March 2003 > Whenever you ask for a sibling instance of a Player in an EToy, the > Players of the new object is an instance of the original Player class, > instead of having a new class. When you duplicate a morph the holds a > set of siblings, there was bug that made the new objects not be > siblings. This fixes the bug. Alan wants this fixed for the cells of > a Tic-Tac-Toe game. > To test, make a playfield with a blob in it. Make a sibling of the > blob. Tell the playfield to 'make sibling instance'. All four blobs > will remain siblings. > In addition, the production of a sibling instance is now cleaner. > (Updated to fix a conflict with the 5240 MCP update. -dew)" > > 5247VeryDeepCopyUsing-tk -- Ted Kaehler -- 13 May 2003 > Object>>veryDeepCopyUsing: is a utility method. When you have a > complex data structure, involving several custom classes, veryDeepCopy > on a subpart may not do what you want. When you veryDeepCopy one of > your objects, it does not know which instance variable are actually > pointers back to a 'global' object, which should not be duplicated. > The 'parent' links in a tree are an example. These can be handled by > creating veryDeepInner: and veryDeepFixupWith:. However, it may be > easier to make a custom veryDeepCopy method. In it, you create a > DeepCopier, and store into its references dictionary with (globalObject > -> globalObject). Then you want to start a veryDeepCopy using that > DeepCopier, and calling veryDeepCopyUsing: is the way to do that."

1 0

Re: request for critique
by Bob Arning 02 Jun '03

02 Jun '03

On Mon, 2 Jun 2003 06:04:47 -0500 David Faden <dfaden(a)iastate.edu> wrote: >The code is for a little Mandelbrot set explorer. It seems pretty slow >(compared to similar implementations of the M. set in different >environments and compared to other graphics objects within the Squeak >image). Is the usual way to create an image pixel-by-pixel in Squeak to >create a Form then set the pixels one-by-one using #colorAt: put: or >#pixelValueAt: put:? What could I do to improve the performance? David, You seem to be doing the traditional algorithm correctly in Squeak. One thing that speeds things up a bit is a different algorithm. http://www.charm.net/~arning/MandelBob.17Jun431pm.cs Implements something called the Mariani/Silver algorithm which I lifted from a sample application written in MacApp. This approach examines the edges of rectangle and, when all edge pixels are the same, fills the interior without checking each pixel. I don't know if this has been proven to be correct, but it certainly looks right to the eye and is a good bit faster. >Is there a way to associate an hour glass cursor with a single Morph to >allow it indicate that it's carrying out a noticeably time consuming >process? It isn't so much associating an hour glass with a morph as it is showing the hour glass while some code is running. The way to do this is... Cursor wait showWhile: [your code in here]. Alternatively, you can move the computation to another process and left the user have the use of the computer while it's running. That's the approach I took. If you decide to file in the above code to a recent Squeak image, at leadt one thing needs to change... !MandelMorph2 methodsFor: 'menu' stamp: 'RAA 6/2/2003 08:34'! mouseDown: evt (MenuMorph new) add: 'Zoom in' target: self selector: #zoomIn; "--" add: 'Save to disk' target: self selector: #saveToDisk; "--" add: 'Set project background' target: self selector: #setProjectBackground; "--" popUpAt: Sensor cursorPoint forHand: evt hand in: World. ! ! Cheers, Bob

1 0

RE: [ENH] CelesteShowMoveAgainTarget
by Frank Sergeant 02 Jun '03

02 Jun '03

"Lex Spoon" <lex(a)cc.gatech.edu> wrote: > I believe the easist thing would be to add a classic UI to Filtering > Celeste. FC has (or at least *had*) various goodies over the existing ... > On the other hand, I like FC way better, even with its nearly complete > lack of maintenance. Lex, do you mean you prefer the UI of FilteringCeleste over the UI of the Celeste that is currently in the image? Could you describe the differences and what's better about each? (Should we consider using FilteringCeleste without adding a classic UI to it?) I've been using and enjoying the in-image version of Celeste for some time now as my mail program and have been hoping to try FilteringCeleste but inertia has gotten in the way. I can't (comfortably) browse several of my mail folders now because they contain huge numbers of messages and thus are terribly slow. At least the messages are filed away and one day, with FilteringCeleste and/or the large list stuff, ... -- Frank

2 1

Student Volunteer Program for ESUG'2003
by Stephane Ducasse 02 Jun '03

02 Jun '03

Please distribute: This year again ESUG http://esug2003.esug.org/ is organizing a student volunteer program! You are a student and you want to attend ESUG (the first European Conference on Smalltalk)? ESUG has again a student volunteer program so you can get the conference for free (http://esug2003.esug.org/registration.htm). Your duties will be low and you will have to help a bit the local organizers. ESUG will not pay the travel but the conference will be free and possibly the hosting will be also free depending on the number of students. For the students that cannot afford the traveling cost we are setting up a *private* sponsoring program, some companies are ready to sponsor some students. But pay attention the places are limited and do not wait the last minute to register. Student volunteers will be notified the 14 of July. To register to the student volunteer program please send an email to nano(a)essi.fr **and** ducasse(a)iam.unibe.ch. Please explain **clearly** your situation and your motivation (fill the template below) For the other remember that the early registration is the 15 of July Stef and Olivier Name: Email where you can **always be reached**: Level in Curriculum: Motivation: Explanation why you need the help of ESUG: Special information: ====================================================================== The European Smalltalk User Group invites you to the Smalltalk Joint Event 11th ESUG Smalltalk Conference & ESUG Research Track Tuesday, August 25, - Friday, August 29, 2003 & Camp Smalltalk Saturday, August 23 - Sunday, August 24, 2001 Bled, Slovenia http://esug2003.esug.org/ ======================================================================== GENERAL INFORMATION =================== For the 11th consecutive year, the European Smalltalk User Group organizes a Smalltalk Conference to bring together Smalltalk users from both academia and industry in a friendly atmosphere. As always the attendants will enjoy tutorials, workshops and demonstrations on various Smalltalk related subjects. New this year is that the talks are scheduled primarily in the morning and that the afternoon focuses on concrete collaborations between participants and workshops. Continuing a strong tradition in organizing Smalltalk events for Smalltalk practitioners, this year promises to be very exiting. ESUG organizes a Smalltalk Joint Event comprised of Camp Smalltalk, Advanced Seminars and a Research track: - An official (and free) Camp Smalltalk, starting the week-end before and continuing the rest of the week. - The Advanced Seminars, where invited speakers and technical presenters show concrete applications and research using Smalltalk. - ESUG Research Track: an academic research track with top-of-the-line program committee. A separate call for paper was issued for this event, which is attached to this mail at the bottom. All these exiting events will take place in the last week of August, between the 23rd and 29th. As always, we cannot make this work without contributions from the actual Smalltalk practitioners. Therefore we invite people to submit proposals to present papers (both academic or industrial, technical or experience report), or to exhibit. Together with invited speakers and Campers this promises to be another exciting Smalltalk event. INSTRUCTIONS FOR PROSPECTIVE SPEAKERS ===================================== Prospective speakers must clearly explain the format they would like to use (tutorial, experience report, workshop or demo). Each proposal must include a valid e-mail address and an abstract of around 200 words. All proposals have to be sent to Roel Wuyts (wuyts(a)iam.unibe.ch) AND Stephane Ducasse (ducasse(a)iam.unibe.ch), in plain ASCII pasted in the body of the e-mail (NOT attached). Hot topics include, but are not limited to, - web service applications using Smalltalk - experience using eXtreme Programming - distributed applications - new Smalltalk implementations and optimizations - Smalltalk Development Tools - Language improvements or enhancements Exhibitors or sponsors should directly contact the ESUG board for more information (esug-info(a)esug.org). PRACTICAL INFORMATION ===================== The Conference takes place in Bled, in Slovenia. Bled is a city in the north west of Slovenia, under mountains. It is near the Austrian and the Italian border, with the beautiful Italian city of Venice only three hours away. More information about the city of Bled can be found in English at the following URL: http://www.bled.si/eng/informacije/obledu.htm. The conference will be held in hotel Golf (http://www.gp-hoteli-bled.si/ang/golf.htm). This hotel has all necessary facilities (internet and beamers) to support a successful organization. Conference attendees have to book accommodation themselves this year, either at this hotel or at other hotels in Bled. CONFERENCE FEES =============== http://esug2003.esug.org/registration.htm Early (until july 15) Late (after july 15) Academic 220 EUR 399 EUR Industry 385 EUR 499 EUR IMPORTANT DATES =============== deadline for proposals: April 11th (But we are always looking for excellent talks). camera ready copy for slides of accepted speakers: July 7th ======================================================================== CALL FOR PAPERS ESUG Research Track 25-29 Aug 2003 Bled Slovenia http://www.esug.org/ ======================================================================== The European Smalltalk User Group is proud to announce that the 11th Annual ESUG Conference will include a refereed academic track. A strong program committee has been selected and we expect the academic track to include high quality papers. Proceedings will be published. Scope: ====== The goal of the academic track is to have a forum for academic publications related to research in Smalltalk and dynamically typed languages. We encourage authors to submit excellent quality papers as we plan to produce proceedings. A non exhaustive list of topics is - new languages features (mixins, AOP,...) - multi-agent systems - meta and reflective programming - code analysis (refactoring,...) - process development (Agile processes, Unit testing) - virtual machines (optimization, new trends) - integrated development environments (browsers, visualization, ...) - frameworks (web, graphical...) - software evolution (metrics,...) Program Chair: ============== Stephane Ducasse (University of Berne) email: ducasse(a)iam.unibe.ch Roel Wuyts (University of Berne) email: roel.wuyts(a)iam.unibe.ch Program Committee: ================== - Prof. Andrew Black (Oregon Health and Science University, USA) - Dr. Noury Bouraqadi (Ecole de Mines de Douai, France) - Prof. Serge Demeyer (Universiteit Antwerpen, Belgium) - Prof. Theo D'Hondt (Universiteit Vrije Brussels, Belgium) - Prof. Stephane Ducasse (Universitaet Bern, Switzerland) - Prof. Ralph Johnson (University of Illinois, USA) - Eliot Miranda (Cincom, USA) - Joseph Pelrine (MetaProg, Switzerland) - Dave Simmons (Quasar, USA) - Dr. Serge Stinckwich (University of Caen, France) - Dave Thomas (Bedarra, USA-Canada) - Dr. Roel Wuyts (Universitaet Bern, Switzerland) Important Dates: ================ - Deadline: 15th of April 2003 (and not 15th of March as originally mentioned) - Notification of acceptance: 1st of June 2003 - Final version: 1st of July 2003 Important Information: ====================== - Best papers will be published in a special issue of the journal of Computer Languages - Preferred format: PDF - Maximum paper length 15 pages Information related to the ESUG Academic Track inside the ESUG Conference Organization: ======================================================================== We plan to have the Academic track the first day of the Conference: the 25 of August 2003.

1 0

[KCP] is looking for reviewers
by Stephane Ducasse 02 Jun '03

02 Jun '03

Hi guys we need a bit of pressure to make progress. A lot of good cleans are waiting to be reviewed and approved. The more we wait the higher the probability of conflicts. http://minnow.cc.gatech.edu/squeak/3083 (we will clean the page soon) Stef

1 0

[UPDATES] 7 for 3.6alpha
by Doug Way 02 Jun '03

02 Jun '03

These are the 4 items which were [approved] last week, plus one which Daniel approved about a month ago that I somehow missed. (The subject line for that one was munged a bit, though... I'll use that as an excuse. ;-) ) - Doug Way ------------------------------- 5241instrPrinter-md -- Marcus Denker -- 8 April 2003 A refactoring for InstructionPrinter, AbstractInstructionPrinter and InstVarRefLocator: AbstractInstructionPrinter: moved the instVar and a method into InstVarRefLocator, now this class can be used as a Abstract Superclass for both InstructionPrinter and InstVarRefLocator Renamed to InstructionClient (c.f. VisualWorks). InstructionPrinter: refactored to be a subclass of InstructionClient" 5242HashChanges1-SqR -- Andres Valloud, Stephan Rudlof -- 13 August 2002 Goals: Implement an improved hash for large integers. Refactor/rename some hash methods to make them more intention revealing. Pass 1 - Implement refactoring so it can be bridged later. sr: I have changed #byteAt: to #basicAt: to make it Slang compilable (and changed some var names)." 5243HashChanges2-SqR -- Andres Valloud -- 13 August 2002 Pass 2. Implement LargePositiveInteger>>hash. Rehash sets as appropriate." 5244HashChanges3-SqR -- Andres Valloud -- 13 August 2002 Pass 3. Make String and ByteArray use the refactoring instead of the original implementation." 5245MorphicClsComments-efc -- Eddie Cottongim -- 7 March 2003 Class Comments for Morph, ImageMorph, SimpleButtonMorph and StringMorph. These were subject to review and editing on the swiki, and thus should be fairly comprehensive." 5246Sibling-dup-tk -- Ted Kaehler -- 11 March 2003 Whenever you ask for a sibling instance of a Player in an EToy, the Players of the new object is an instance of the original Player class, instead of having a new class. When you duplicate a morph the holds a set of siblings, there was bug that made the new objects not be siblings. This fixes the bug. Alan wants this fixed for the cells of a Tic-Tac-Toe game. To test, make a playfield with a blob in it. Make a sibling of the blob. Tell the playfield to 'make sibling instance'. All four blobs will remain siblings. In addition, the production of a sibling instance is now cleaner. (Updated to fix a conflict with the 5240 MCP update. -dew)" 5247VeryDeepCopyUsing-tk -- Ted Kaehler -- 13 May 2003 Object>>veryDeepCopyUsing: is a utility method. When you have a complex data structure, involving several custom classes, veryDeepCopy on a subpart may not do what you want. When you veryDeepCopy one of your objects, it does not know which instance variable are actually pointers back to a 'global' object, which should not be duplicated. The 'parent' links in a tree are an example. These can be handled by creating veryDeepInner: and veryDeepFixupWith:. However, it may be easier to make a custom veryDeepCopy method. In it, you create a DeepCopier, and store into its references dictionary with (globalObject -> globalObject). Then you want to start a veryDeepCopy using that DeepCopier, and calling veryDeepCopyUsing: is the way to do that."

1 0

[BUG] SAR/KCP conflict
by Doug Way 02 Jun '03

02 Jun '03

This is a fix for the SAR package, whose method PositionableStream>>fileInFor:announcing: still used an old SystemDictionary method instead of the new SystemNavigation methods. This prevented the SUnit package from being installed into a current 3.6alpha-5240 image, for example. Hmm, I guess if this change is made to SAR, then it won't work with Squeak 3.4, which will prevent people from installing the SqueakMap bootstrap for 3.4. Perhaps a more defensive test needs to be put in for the time being... maybe test if the SystemNavigation class is present and use that if it is, otherwise use the Smalltalk global variable. Argh. So, this isn't quite a complete fix. Goran, we need SM 1.1 pretty soon so that we can have different versions of these things... :-) - Doug Way

2 3

Callbacks and FFI
by Emeric Rollo 02 Jun '03

02 Jun '03

Hi all! I'd like to use a C library in squeak. (using FFI) But I saw on squeak-dev's archives that callbacks functions are not supported by FFI stuff... However, these messages are old (2000-2001). Does anyone know if there is currently a way to call callback functions with FFI ? Thanks. Emeric.

2 1

StrikeFont(NewYork12 15) becomes bold after #saveOnFile
by Stephane Rollandin 02 Jun '03

02 Jun '03

hello list, I'm submitting again this problem I've had for a while since my previous post (a few weeks ago) did not get any answer. here we go: ---- my default font is NewYork12 15 when I save a TextMorph or a MenuItemMorph with #saveOnFile, then import the .morph file back in the image, the font becomes NewYork12B 15, bold. OTOH there is no problem with NewYork10 12 (the default one in Squeak 3.4). no problem with StringMorph either. how could I fix this ? Stef

4 6

Peripherals 2
by Alistair Wallis 01 Jun '03

01 Jun '03

Thanks for advice that has been given. I was hoping to programme a LEGO RCX. Does anyone know how this can be done?

5 5

Jump to page: